US9148620B2 - Detecting video formats - Google Patents

Detecting video formats Download PDF

Info

Publication number
US9148620B2
US9148620B2 US13/083,681 US201113083681A US9148620B2 US 9148620 B2 US9148620 B2 US 9148620B2 US 201113083681 A US201113083681 A US 201113083681A US 9148620 B2 US9148620 B2 US 9148620B2
Authority
US
United States
Prior art keywords
cadence
variance
current
video
difference
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US13/083,681
Other versions
US20120257104A1 (en
Inventor
Jong Dae Oh
Yi-Jen Chiu
Sang-Hee Lee
Ning Lu
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Intel Corp
Original Assignee
Intel Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Intel Corp filed Critical Intel Corp
Priority to US13/083,681 priority Critical patent/US9148620B2/en
Assigned to INTEL CORPORATION reassignment INTEL CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: LU, NING, OH, JONG DAE, CHIU, YI-JEN, LEE, SANG-HEE
Publication of US20120257104A1 publication Critical patent/US20120257104A1/en
Application granted granted Critical
Publication of US9148620B2 publication Critical patent/US9148620B2/en
Application status is Active legal-status Critical
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00Television systems
    • H04N7/01Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
    • H04N7/0112Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level one of the standards corresponding to a cinematograph film standard
    • H04N7/0115Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level one of the standards corresponding to a cinematograph film standard with details on the detection of a particular field or frame pattern in the incoming video signal, e.g. 3:2 pull-down pattern

Abstract

The format of telecined video may be determined including a bottom field first cadence. In addition, video using 2:3:3:2 top field first can be identified. Moreover, mixed cadence videos can also be detected. In some embodiments, mixed cadence videos may be detected by calculating variances of different areas within a frame.

Description

BACKGROUND

This relates generally to processing of video information and, particularly, to identifying video format.

Video may have different formats based on how it is initially recorded. Different recording or imaging devices may record the video in different formats. As an example, different video formats may include interlaced versus progressive formats, using top field first or bottom field first, etc.

In order to process video, it is desirable to know which format the video is actually in. Knowing this information, the format may be changed to better match the format generally used by an available video display device. If the video is not formatted properly for the display, viewers may see interlacing artifacts in the displayed images.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flow chart for one embodiment of the present invention; and

FIG. 2 is a hardware schematic depiction of one embodiment of the present invention.

DETAILED DESCRIPTION

In accordance with some embodiments, the format of telecined video may be determined automatically and may be reconstructed into another format.

Video can be broken into two fields. In interlaced video, one line from each field is displayed alternately. In progressive video, the two fields are not interlaced. In top field first video, the top field is displayed first and, in bottom field first, the bottom field is displayed first or at the top.

Telecined video generally has six different cadences. In addition there are two special cases, mixed cadence video and progressive versus interlaced video. Mixed cadence video is video which has two different cadences in the same display or frame. Progressive versus interlaced video detection are two alternate video formats.

In accordance with some embodiments, 16 variances of two consecutive frames may be calculated. There are three major variance records or variables that arise from the variances. One major variance is the top field repetition between the current and previous frames. For example, the top field may be the same or it may be different between two particular consecutive frames. Another major variance is bottom field repetition between current and previous frames. The third major variance is interlaced video information that is based on the difference between the current top field and bottom field.

Based on these three variables or variance records, a confidence level may be determined for the current cadence or video format. If the confidence level is high enough, the current cadence may be confirmed first. Otherwise, the system may try to lock and detect a new cadence.

The 16 video formats or cadences are set forth in the chart below.

TABLE 1
Cadences Periodicity Frame patterns
60i 1 T0B1 T2B3 T4B5 T6B7 T8B9 . . .
30 fps FRP: 00 00 00 00 00 . . .
ITL: 1 1 1 1 1 . . .
PFF: 00 00 00 00 00 . . .
2:2 (30p) 1 T0B0 T1B1 T2B2 T3B3 T4B4 . . .
30 fps FRP: 00 00 00 00 00 . . .
ITL: 0 0 0 0 0 . . .
PFF: 00 00 00 00 00 . . .
3:2 (BFF) 5 T0B0 T1B0 T2B1 T2B2 T3B3 T4B4 T5B4 T6B5 T6B6 T7B7
(OddFF) FRP: 00 01 00 10 00 00 01 00 10 00 . . .
24 fps ITL: 0 1 1 0 0 0 1 1 0 0 . . .
PFF: 00 01 01 01 00 00 01 01 01 00 . . .
2:3 (TFF) 5 T0B0 T0B1 T1B2 T2B2 T3B3 T4B4 T4B5 T5B6 T6B6 T7B7
(EvenFF) FRP: 00 10 00 01 00 00 10 00 01 00 . . .
24 fps ITL: 0 1 1 0 0 0 1 1 0 0 . . .
PFF: 0 10 10 10 00 00 10 10 10 00 . . .
2:3:3:2 5 T0B0 T1B0 T2B1 T3B2 T3B3 T4B4 T5B4 T6B5 T7B6 T7B7
(BFF) FRP: 00 01 00 00 10 00 01 00 00 10 . . .
24 fps ITL: 0 1 1 1 0 0 1 1 1 0 . . .
PFF: 00 01 01 01 01 00 01 01 01 01 . . .
2:3:3:2 5 T0B0 T1B1 T1B2 T2B2 T3B3 T4B4 T5B5 T5B6 T6B6 T7B7
(TFF) FRP: 00 00 10 01 10 00 00 10 01 00 . . .
24 fps ITL: 0 0 1 0 0 0 0 1 0 0 . . .
PFF: 00 00 10 10 00 00 00 10 10 00 . . .
2:2:2:4 5 T0B0 T1B1 T2B2 T3B3 T3B3 T4B4 T5B5 T6B6 T7B7 T7B7
24 fps FRP: 00 00 00 00 11 00 00 00 00 11 . . .
ITL: 0 0 0 0 0 0 0 0 0 0 . . .
PFF: 00 00 00 00 11 00 00 00 00 11 . . .
3:2:3:2:2 6 T0B0 T1B0 T2B1 T2B2 T3B3 T4B4 T5B5 T6B5 T7B6 T7B7
(BFF) FRP: 00 10 00 01 00 00 00 10 00 01 00 00 . . .
24 fps ITL: 0 1 1 0 0 0 0 1 1 0 0 0 . . .
PFF: 00 10 01 10 00 00 00 10 01 10 00 00 . . .
3:2:3:2:2 6 T0B0 T0B1 T1B2 T2B2 T3B3 T4B4 T5B5 T5B6 T6B7 T7B7
(TFF) FRP: 00 00 01 00 10 00 00 00 01 00 10 00 . . .
24 fps ITL: 0 0 1 1 0 0 0 0 1 1 0 0 . . .
PFF: 00 00 01 01 01 00 00 00 01 01 01 00 . . .
6:4 5 T0B0 T0B0 T0B0 T1B1 T1B1 T2B2 T2B2 T3B3 T3B3 T3B3
12 fps FRP: 00 11 11 00 11 00 11 11 00 11 . . .
ITL: 0 0 0 0 0 0 0 0 0 0 . . .
PFF: 00 11 11 00 11 00 11 11 00 11 . . .
5:5 5 T0B0 T0B0 T1B0 T1B1 T1B1 T2B2 T2B2 T3B2 T3B3 T3B3
(BFF) FRP: 00 11 01 10 11 00 11 01 10 11 . . .
12 fps ITL: 0 0 1 0 0 0 0 1 0 0 . . .
PFF: 00 11 01 01 11 00 11 01 01 11 . . .
5:5 5 T0B0 T0B0 T0B1 T1B1 T1B1 T2B2 T2B2 T2B3 T3B3 T3B3
(TFF) FRP: 00 11 10 01 10 00 11 10 01 11 . . .
12 fps ITL: 0 0 1 0 0 0 0 1 0 0 . . .
PFF: 00 11 10 10 11 00 11 10 10 11 . . .
8:7 15 T0B0 T0B0 T0B0 T0B0 T1B1 T1B1 T1B1 T2B1 T2B2
(BFF) T2B2 T2B2 T3B2 T3B3 T3B3 T3B3 T4B4 T4B4 . . .
8 fps FRP: 00 11 11 11 00 11 11 01 10
00 00 10 01 00 00 00 11 . . .
ITL: 0 0 0 0 0 0 0 1 0
0 0 1 0 0 0 0 0 . . .
PFF: 00 11 11 11 00 11 01 10 11
11 01 10 11 11 00 11 . . .
8:7 15 T0B0 T0B0 T0B0 T0B0 T1B1 T1B1 T1B1 T1B2 T2B2
(tFF) T2B2 T2B2 T2B3 T3B3 T3B3 T3B3 T4B4 T4B4 . . .
8 fps FRP: 00 11 11 11 00 11 11 01 10
00 00 10 01 00 00 00 11 . . .
ITL: 0 0 0 0 0 0 0 1 0
0 0 1 0 0 0 0 0 . . .
PFF: 00 11 11 11 00 11 11 01 10
11 11 01 10 11 11 00 11 . . .
2:3:3:2-1 5 T0B0 T1B1 T2B2 T3B2 T3B3 T4B4 T5B5 T6B6 T7B6 T7B7
(BFF) FRP: 00 00 00 10 11 00 00 00 10 11 . . .
5 fps ITL: 0 0 0 1 0 0 0 0 1 0 . . .
PFF: 00 00 00 01 00 00 00 00 01 00 . . .

Taking as an example the 2:3 (top field first TFF) (even fields) at 24 frames per second (fps), the first frame (T0B0) is made up of a top field zero (T0) and a bottom field zero (B0). Thus, this is an example of top field first video format or cadence.

The next frame (T0B1) is the top field zero (T0), having been repeated, and a new bottom field, the bottom field number one (B1). Thus, in Table 1, frame T0B0 is followed by frame T0B1. The next line in the chart gives the field repetition pattern (FRP). The field repetition pattern for the first frame T0B0 is made up of two values, both of which are zero to indicate that there has not yet been any field repetition. But then examining T0B0 versus the next frame T0B1, the top field has been repeated since the top field is T0 in both cases. Thus, the first bit of the frame repeat pattern is one (after the second frame T0B1), to indicate that there is a repetition in the top field and the second bit is zero, to indicate that there is no repetition in the bottom field.

The next line is for interlaced video (ITL). Zero indicates that the frame is not interlaced and one indicates that it is interlaced.

The final line is the possible field first (PFF), which indicates whether it is a possible top field first or bottom field first. The PFF is initially zero before any frames arrive. The next value in the PEE line is 10. The one bit indicates that the current top field and the previous bottom field fit into a progressive frame and the next bit, which is zero, indicates that the previous top field and the current bottom field do not fit into a progressive frame.

There are also two special detection algorithms for mixed cadence detection and progressive versus interlaced video detection. Usually, at the bottom of the video frame, there is a subtitle for language or other information. Sometimes that subtitle or other information is overlaid on video having a different cadence that the subtitle. In order to detect mixed cadences, two new variances are created. The first new variance is top field difference between the bottom quarter of a current and a previous frame and the other new variance is bottom field difference between the bottom quarter of a current and a previous frame. If there is a mismatched cadence between the whole image and its bottom quarter, the video may be handled by a special routine for mixed cadence video and reconstructed using a de-interlacing engine in the default mode.

For progressive frame detection, motion information for each pixel between the current and previous frame, between top fields in the current frame and between bottom fields in the current frame are evaluated.

In accordance with some embodiments, there are 16 variances that are accounted for, together with a confidence level, mtag variable, and three variance records may be calculated in a first stage. The first 14 (variances [0]-[13]) of the 16 variances are differences for top and bottom fields for an entire current and previous frame. The last two variances (variances [14] and [15]) are for top and bottom field differences for the bottom one quarter of the current and previous frames. The 16 variances are set forth below.

TABLE 2
Variance Description
Variance[0] Number of interlaced high-variance pixel pairs
Variance[1] Number of progressive high-variance pixel pairs
Variance[2] Difference between the previous bottom field and the
current top field (CtPb)
Variance[3] Difference between the previous top field and the current
bottom field (CbPt)
Variance[4] Difference between the previous top field and the current
top field (CtPt)
Variance[5] Difference between the previous bottom field and the
current bottom field (CbPb)
Variance[6] Difference between the current top field and the current
bottom field (CtCb)
Variance[7] total moving pixels
Variance[8] Find tears between current top and bottom fields
Variance[9] variance[7]-variance[8]
Variance[10] Find tears between current top and previous bottom fields
Variance[11] variance[7]-variance[10]
Variance[12] Find tears between current bottom and previous top fields
Variance[13] variance[7]-variance[12]
Variance[14] CtPt for ¾ * height to height
Variance[15] CbPb from ¾ * height to height

The first two variances (variances [0] and [1]) are based on the number of high-variance pixel pairs. They are identified and counted. First, a check determines whether the sum of the difference between the current and previous frame is larger than a temporal threshold, and whether current top and current bottom fields are different enough. If so, the interlaced high variance pixel is increased by 1.

The pseudocode is as follows:

if(Difference between Current top (cT) and
Previous top (pT) plus difference between Current
bottom (cB) fields and Previous bottom field (pB) >
Temporal Threshold)
if(Difference_cT_cB > Difference_cT_cT +
Difference_cB_cB) Increase the number of
interlaced high-variance pixel pairs by 1

Otherwise, if the current top and current bottom fields are riot different enough, the number of progressive high-variance pixel pairs is increased by 1.

Variances [2] through [6] are based on field differences. Assume that a current pixel is at (x,y) and y is the current top field line. The difference_cT_pT is equal to the square of (C(x,y)−P(x,y)), the difference_cB_pB=square of (C(x,y+1)−P(x,y+1)), the difference_cT_cB=square of (C(x,y)−C(x,y+1)), the difference_cT_cT=square of (C(x,y)−C(x,y+2)), and the difference_cB_cB=square of(C(x,y−1)−C(x,y+1).

Variance 7 is based on the total moving pixels. The difference_cT=ABS(C(x,y−2)−C(x,y))+ABS(C(x,y)−C(x,y+2))+ABS(C(x,y−2)−C(x,y+2)) and difference_cB=ABS(C(x,y−1)−C(x,y+1))+ABS(C(x,y+1)−C(x,y+3))+ABS(C(x,y−1)−C(x,y+3)). Therefore, if the difference between the current frame and the previous frame is big enough (which means two co-located pixels are very different), and the difference between current top field and difference between bottom field are small enough (which means the pixel area does not have high texture areas and it would be difficult to get a big difference between the current and previous frames), then the number of moving pixels is increased by 1.

The pseudocode for determining the total moving pixels is as follows:

if(Difference_cT < Vertical Threshold &&
Difference_cB < Vertical Threshold)
Increase the number of moving pixels by 1.

Variances [8] through [13] are based on the existence of tears. There are three kinds of tears. If a given pixel is a moving pixel, the difference between the given two top and bottom fields is above a threshold, and the difference of the two different fields is larger than the sum of two same field difference, the tear pixel number is increased. So, a tear uses the same idea as the interlaced high variance pixel with a different combination of top and bottom fields.

The pseudocode is as follows:

if(Difference_cT_cB >= Vertical Threshold &&
Difference_cT_cB > Difference_cT_cT +
Difference_cB_cB)  Increase Tear pixel_0 by 1;
  if(Difference_cT_pB >= Vertical
Threshold && Difference_cT_pB > Difference_cT_cT +
Difference_pB_pB)  Increase Tear pixel_1 by 1;
  if(Difference_cB_pT >= Vertical
Threshold && Difference_cB_pT > Difference_cB_cB +
Difference_pT_pT)  Increase Tear pixel_2 by 1;
}

The variable mtag represents a relationship between variances [4] and [5]. By normalizing and comparing variances [4] and [5] to a threshold, top and bottom field repetition can be calculated. When variance [4] is much bigger than variance [5], then the variable may be set to the number 4. When the variance [5] is much bigger than the variance [4], then the variable is 1 and, otherwise, the variable is 0. The variable updates an mtag history for fast cadence locking. This technique can be implemented by the following equation:
if ((variance[4]/α)>variance[5])
mtag=4;
else if ((variance[5]/α)>variance[4])
mtag=1;
else
mtag=0;
where alpha is a threshold to determine the amount of difference between variance [4] and variance [5].

Thus, referring to FIG. 1, the sequence 10 may be implemented in software, firmware, or hardware. In a software embodiment, it may be implemented by instructions stored in a non-transitory computer readable medium, such as a magnetic, semiconductor, or optical memory. The sequence of instructions may be implemented by a processor or computer.

Once the variances are calculated (block 12), a check at diamond 14 determines whether the confidence level for the current case is sufficient. If the current confidence level is not high enough then a confirming cadence is determined at diamond 20. Depending on the three variance records, the cadence may be confirmed in diamond 20, based on whether the current cadence is correct or not. The three variance records may be compared to the fixed pattern for cadence in Table 1. If the current variable records are not identical to a fixed pattern, then the flow proceeds to checking cadence in diamond 22. Otherwise, a confidence value may be increased and the flow proceeds to checking the scene change in block 24.

In diamond 22, the variable records are calculated with new thresholds. The first step of this function is to calculate two thresholds that determine high and low variance values. Then the current cadence may be combined with the calculated page based on the two thresholds. The page is equal to the current frame number modulo the periodicity of the current cadence. The last step of checking cadence determines whether the three variance records of the estimated page are consistent with the fixed records of the current mode. If there is a match between the estimated page and the fixed current value, the flow proceeds to scene change check at block 24. Otherwise, the flow proceeds to detecting cadence at diamond 18 in order to estimate a new cadence. Detecting cadence is the same function used after locking cadence in diamond 16, which is a function for low confidence.

The locking cadence function 16 compares the three variant recodes with a fixed pattern of cadence. If there is a matched cadence, the new cadence is updated with the new page and confidence. If the newly estimated mode is not consistent with a previous cadence, then the flow goes to detecting cadence at diamond 18. The detecting cadence function may include two steps. In a first step, with the three variance records, new cadence, page, and confidence are estimated for modes 5:5, 2:3:3:2, 3:2:3:2:2, and 8:7. If there is no match, the current first and second variance records are compared for modes 2:3,3:2, and 2:3:3:2. If there is no match with a cadence, the current cadence is set to 60 interlaced (60i), which is the default cadence in one embodiment.

The scene change check at block 24 is a fallback mode for a frame that has a scene change. There are two summation values in the following equation:
FMD_Sums [0]+=variance [4]−variance [4] at previous 15 frames
FMD_Sums[1]+=variance[5]−variance [5]at previous 15 frames

Because there were no previous variances before the first 15 frames, FMD_Sums [0] for the first 15 frames accumulates variance [4], while FMD_Sums [1] for the first 15 frames accumulates variance [5]. After the first 15 frames, the two sums are updated with new variances and the old variance values are subtracted. If variance [4] is larger than half of FMD_Sums [0], a scene change is detected at the current frame and the current cadence is set to 60 interlaced. Also, if variance [5] is larger than half of FMD_Sums[1], a scene change is detected and the cadence is set to 60i.

In the next step (block 26), cadence estimating is based on the mtag and FMD_SeqMode. The 2:3 and 3:2 cadences have a small periodicity which is 5 frames. Therefore, these cadences may be used as a current cadence, while the algorithm goes to the final cadence. The mtag represents the relationship between the top field and the bottom field repetition. The FMD_SeqMode is a record for mtag when the current cadence is either 2:3 or 3:2. If a previous cadence is either 3:2 or 2:3, the last digit of FMD_SeqMode is increased by 1. If the current cadence is either 3:2 or 2:3 and mtag equals 1, the last digit of FMD_SeqMode is set to 2. If the previous mode is either 3:2 or 2:3 and mtag equals 4, the last digit of the FMD_SeqMode is set to 5. So if there are consecutive 2:3 and 3:2 modes of more than two frames in a row, then the cadence can be estimated based on m_FMD_SeqMode. By comparing FMD_SeqMode and the fixed mtag record value, the current cadence can be estimated promptly without confidence checking.

The next step (block 28) may involve special detection for 2:2 versus 60i based on variances [8]-[13]. The variances [8]-[13] represent a tear in a current and previous frames. For example, variance [8] is a number of tear pixels between current top and bottom fields. If the difference between the current top and bottom fields is larger than a sum of differences between the current top field lines at y and the line below the current line by 2 lines (y−2) and a difference between current bottom lines at y−1 and the 3 lines below the current line (y−3), this pixel is denominated a tear pixel. By comparing the number of tear pixels between current top and bottom fields with current top and previous bottom fields, and between current bottom and previous top fields, the video can be identified as progressive or interlaced video.

Once a cadence and page have been estimated for the current frame, the frame restoration (block 30) becomes straightforward. The FMD function can be called for every other field. Therefore, there are two pages. One page is for the previous field and the other page is for the current field. The page of each cadence has information for repeating top and bottom fields or not repeating. Thus, there are three combinations. The top field could be the previous frame and bottom field could be the current frame, the top field could be the current frame and the bottom field could be the previous frame, or the top field could be the current frame and the bottom field is the previous frame.

In accordance with some embodiments, 14 variables and 3 variance records may be used to match the current calculated pattern with a fixed top field first pattern. By confidence checking, the conventional algorithm changes the cadence.

Some embodiments can handle not only top field first, but 7 new cadences as well, 6 for bottom field first and another for 2:3:3:2 of top field first. To detect a cadence quickly, in some embodiments, the record of the relationship between the top field and the bottom field repetition is used. By using the record of that relationship, the system can detect a correct cadence without a confidence check. By combining these two informations, the cadence can be detected quickly and accurately, in some embodiments.

In addition, progressive video may be detected by generating three different tear pieces of information based on the current and previous frames and obtaining the ratios of interlaced to progressive-like pixels. By comparing these ratios based on the characteristics of interlaced and progressive video sequences, progressive frames may be classified for mixed, interlaced/progressive frame sequences.

For more accurate detection of a mixed cadence video, the same types of variances may be calculated in different areas of the frame. By comparing statistics between the whole image and part of the image, mixed cadence can be detected before overlaid text affects statistics for the whole image.

The graphics processing techniques described herein may be implemented in various hardware architectures. For example, graphics functionality may be integrated within a chipset. Alternatively, a discrete graphics processor may be used. As still another embodiment, the graphics functions may be implemented by a general purpose processor, including a multicore processor.

References throughout this specification to “one embodiment” or “an embodiment” mean that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation encompassed within the present invention. Thus, appearances of the phrase “one embodiment” or “in an embodiment” are not necessarily referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be instituted in other suitable forms other than the particular embodiment illustrated and all such forms may be encompassed within the claims of the present application.

While the present invention has been described with respect to a limited number of embodiments, those skilled in the art will appreciate numerous modifications and variations therefrom. It is intended that the appended claims cover all such modifications and variations as fall within the true spirit and scope of this present invention.

Claims (26)

What is claimed is:
1. A method comprising:
identifying mixed cadence videos, wherein mixed cadence videos are videos which have two different cadences in the same display or frame by calculating three variances, the first variance being a difference between top field recognition between a current and previous frame, a second variance between a difference between bottom field repetition between a current and previous frame and a third variance based on a difference between a current top field and bottom field.
2. The method of claim 1 including calculating variances of different areas within a frame.
3. The method of claim 1 including identifying a cadence of video by using the relationship between top field and bottom field repetition.
4. The method of claim 3 including identifying the cadence of video without a confidence check.
5. The method of claim 1 including identifying progressive video based on tear information.
6. The method of claim 5 including determining the ratio of interlaced and progressive-like pixels.
7. The method of claim 1 including determining the total number of moving pixels.
8. The method of claim 1 including determining whether variance records correspond to a known fixed cadence.
9. A non-transitory computer readable medium storing instructions, executed by a processor, to
identify mixed cadence videos, wherein mixed cadence videos are videos which have two different cadences in the same display or frame by calculating three variances, the first variance being a difference between top field recognition between a current and previous frame, a second variance between a difference between bottom field repetition between a current and previous frame and a third variance based on a difference between a current top field and bottom field.
10. The medium of claim 9 further storing instructions to calculate variances of different areas within a frame.
11. The medium of claim 9 further storing instructions to identify a cadence of video by using a relationship between top and bottom field repetition.
12. The medium of claim 11 further storing instructions to identify the cadence of video without a confidence check.
13. The medium of claim 9 further storing instructions to identify progressive video based on tear information.
14. The medium of claim 13 further storing instructions to determine the ratio of interlaced and progressive-like pixels.
15. The medium of claim 9 further storing instructions to determine the total number of moving pixels.
16. The medium of claim 9 further storing instructions to determine whether variance records correspond to a known fixed cadence.
17. An apparatus comprising:
a storage; and
a processor coupled to said storage, said processor to identify a mixed cadence video, wherein mixed cadence videos are videos which have two different cadences in the same display or frame by calculating three variances, the first variance being a difference between top field recognition between a current and previous frame, a second variance between a difference between bottom field repetition between a current and previous frame and a third variance based on a difference between a current top field and bottom field.
18. The apparatus of claim 17 wherein said apparatus is a graphics processing unit.
19. The apparatus of claim 17, said processor to identify video using a 2:3:3:2 top field first and to identify mixed cadence videos, wherein mixed cadence videos are videos which have two different cadences in the same display or frame.
20. The apparatus of claim 19, said processor to calculate variances of different areas within a frame.
21. The apparatus of claim 17, said processor to identify a cadence of video by using the relationship between top field and bottom field repetition.
22. The apparatus of claim 21, said processor to identify the cadence of video without a confidence check.
23. The apparatus of claim 17, said processor to identify progressive video based on tear information.
24. The apparatus of claim 23, said processor to determine the ratio of interlaced and progressive-like pixels.
25. The apparatus of claim 17, said processor to determine the total number of moving pixels.
26. The apparatus of claim 17, said processor to determine whether variance records correspond to a known fixed cadence.
US13/083,681 2011-04-11 2011-04-11 Detecting video formats Active 2034-01-06 US9148620B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US13/083,681 US9148620B2 (en) 2011-04-11 2011-04-11 Detecting video formats

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
US13/083,681 US9148620B2 (en) 2011-04-11 2011-04-11 Detecting video formats
EP20120771784 EP2697966A4 (en) 2011-04-11 2012-04-11 Detecting video formats
PCT/US2012/033062 WO2012142118A2 (en) 2011-04-11 2012-04-11 Detecting video formats
JP2014505244A JP2014515234A (en) 2011-04-11 2012-04-11 Detection of video format
CN201280024512.0A CN103563354B (en) 2011-04-11 2012-04-11 Detection Video format
JP2015217983A JP6133382B2 (en) 2011-04-11 2015-11-05 Detection of video format

Publications (2)

Publication Number Publication Date
US20120257104A1 US20120257104A1 (en) 2012-10-11
US9148620B2 true US9148620B2 (en) 2015-09-29

Family

ID=46965837

Family Applications (1)

Application Number Title Priority Date Filing Date
US13/083,681 Active 2034-01-06 US9148620B2 (en) 2011-04-11 2011-04-11 Detecting video formats

Country Status (5)

Country Link
US (1) US9148620B2 (en)
EP (1) EP2697966A4 (en)
JP (2) JP2014515234A (en)
CN (1) CN103563354B (en)
WO (1) WO2012142118A2 (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9496922B2 (en) 2014-04-21 2016-11-15 Sony Corporation Presentation of content on companion display device based on content presented on primary display device
US9940896B2 (en) * 2015-03-03 2018-04-10 Apple Inc. Telecine judder removal systems and methods

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050231596A1 (en) * 2004-03-31 2005-10-20 Marchese Joseph R Testing apparatus for digital video camera anomalies
US7280155B2 (en) 2002-11-06 2007-10-09 Broadcom Corporation Method and system for converting interlaced formatted video to progressive scan video
US7349029B1 (en) * 2005-01-19 2008-03-25 Kolorific, Inc. Method and apparatus for de-interlacing interlaced video fields originating from a progressive video source
US20080158350A1 (en) 2006-12-27 2008-07-03 Ning Lu Method and sytem for telecine detection and restoration
US20080175473A1 (en) * 2007-01-19 2008-07-24 Samsung Electronics Co., Ltd. Device and method of estimating image signal noise and apparatus and method of converting image signal using the same
US20090051809A1 (en) 2007-08-20 2009-02-26 Marshall Charles Capps System and Method for Combining Interlaced Video Frames
US20100225823A1 (en) * 2009-03-06 2010-09-09 Snell Limited Regional film cadence detection
US8027382B2 (en) 2006-06-27 2011-09-27 Apple Inc. Pulldown correction for progressive display of audiovisual recordings
US20120176538A1 (en) * 2011-01-12 2012-07-12 Shekhar Madnani System and method for detecting field order of video sequences

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5406333A (en) * 1994-03-14 1995-04-11 Thomson Consumer Electronics, Inc. Method and device for film-mode detection
JP2003508941A (en) * 1999-08-03 2003-03-04 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ Encoding method and apparatus of a sequence of frames including a video type picture or film type image
WO2003039148A1 (en) * 2001-11-02 2003-05-08 Matsushita Electric Industrial Co., Ltd. Scan conversion apparatus
JP4208590B2 (en) * 2003-02-07 2009-01-14 キヤノン株式会社 2-2 pull-down detection apparatus
KR100518580B1 (en) * 2003-06-10 2005-10-04 삼성전자주식회사 Apparatus and method for performing an inverse telecine process
KR100553893B1 (en) * 2003-10-14 2006-02-24 삼성전자주식회사 Method and apparatus for converting frame rate using time shifting and motion compensation
EP1592246A1 (en) * 2004-04-30 2005-11-02 Matsushita Electric Industrial Co., Ltd. Film mode determination in video image still areas
EP1592247A1 (en) * 2004-04-30 2005-11-02 Matsushita Electric Industrial Co., Ltd. Block mode adaptive motion compensation
US7391468B2 (en) * 2004-07-06 2008-06-24 Magnum Semiconductor, Inc. Telecine conversion detection for progressive scan playback
JP2007300395A (en) * 2006-04-28 2007-11-15 Matsushita Electric Ind Co Ltd Video stream decoding output device
JP2008005183A (en) * 2006-06-22 2008-01-10 Matsushita Electric Ind Co Ltd Video image coding method and coding device suitable for movie material
CN101573962A (en) * 2007-01-03 2009-11-04 皇家飞利浦电子股份有限公司 Film cadence detection
EP1990991A1 (en) * 2007-05-09 2008-11-12 British Telecommunications Public Limited Company Video signal analysis
EP2415259B1 (en) * 2009-04-01 2015-09-16 Marvell World Trade Ltd. Cadence detection in progressive video
US8248527B2 (en) * 2009-08-20 2012-08-21 Broadcom Corporation Method and system for video overlay on film detection on progressive video input

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7280155B2 (en) 2002-11-06 2007-10-09 Broadcom Corporation Method and system for converting interlaced formatted video to progressive scan video
US20050231596A1 (en) * 2004-03-31 2005-10-20 Marchese Joseph R Testing apparatus for digital video camera anomalies
US7349029B1 (en) * 2005-01-19 2008-03-25 Kolorific, Inc. Method and apparatus for de-interlacing interlaced video fields originating from a progressive video source
US8027382B2 (en) 2006-06-27 2011-09-27 Apple Inc. Pulldown correction for progressive display of audiovisual recordings
US20080158350A1 (en) 2006-12-27 2008-07-03 Ning Lu Method and sytem for telecine detection and restoration
US20080175473A1 (en) * 2007-01-19 2008-07-24 Samsung Electronics Co., Ltd. Device and method of estimating image signal noise and apparatus and method of converting image signal using the same
US20090051809A1 (en) 2007-08-20 2009-02-26 Marshall Charles Capps System and Method for Combining Interlaced Video Frames
US20100225823A1 (en) * 2009-03-06 2010-09-09 Snell Limited Regional film cadence detection
US20120176538A1 (en) * 2011-01-12 2012-07-12 Shekhar Madnani System and method for detecting field order of video sequences

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
PCT International Search Report and Written Opinion issued in corresponding PCT/US2012/033062, dated Oct. 29, 2012 (8 pages).

Also Published As

Publication number Publication date
CN103563354A (en) 2014-02-05
JP2016026460A (en) 2016-02-12
US20120257104A1 (en) 2012-10-11
EP2697966A2 (en) 2014-02-19
WO2012142118A3 (en) 2013-03-14
CN103563354B (en) 2017-06-06
WO2012142118A2 (en) 2012-10-18
EP2697966A4 (en) 2014-10-29
JP6133382B2 (en) 2017-05-24
JP2014515234A (en) 2014-06-26

Similar Documents

Publication Publication Date Title
US6473459B1 (en) Scene change detector
US7075581B1 (en) Interlaced-to-progressive scan conversion based on film source detection
US6985187B2 (en) Motion-adaptive interpolation apparatus and method thereof
US8422795B2 (en) Quality evaluation of sequences of images
US7515205B1 (en) Weighted absolute difference based deinterlace method and apparatus
US6757022B2 (en) Method and apparatus for motion adaptive deinterlacing
EP2422526B1 (en) Method and apparatus for automatic transformation of three-dimensional video
US8149336B2 (en) Method for digital noise reduction in low light video
EP1636755B1 (en) Method and system for video quality assessment
US20070223835A1 (en) Block noise detecting method and apparatus, and block noise reducing method and apparatus
US6307888B1 (en) Method for estimating the noise level in a video sequence
US20080062409A1 (en) Image Processing Device for Detecting Chromatic Difference of Magnification from Raw Data, Image Processing Program, and Electronic Camera
CN1694501B (en) Motion estimation employing adaptive spatial update vectors
US20040165084A1 (en) Flicker detecting method and flicker detecting apparatus
US7242819B2 (en) Method and system for advanced edge-adaptive interpolation for interlace-to-progressive conversion
US7170561B2 (en) Method and apparatus for video and image deinterlacing and format conversion
US6563550B1 (en) Detection of progressive frames in a video field sequence
US8446525B2 (en) Edge detection
US9137424B2 (en) Method for flicker detection in image signal
US7259794B2 (en) De-interlacing device and method therefor
JP4901848B2 (en) Video aligning apparatus, video aligning method, video aligning program, and video quality assessing apparatus
US20050232357A1 (en) Motion vector estimation at image borders
CN101836441B (en) Method, apparatus, and system for line-based motion compensation in video image data
KR100957479B1 (en) Method and apparatus for spatio-temporal deinterlacing aided by motion compensation for field-based video
KR100865248B1 (en) Detecting subtitles in a video signal

Legal Events

Date Code Title Description
AS Assignment

Owner name: INTEL CORPORATION, CALIFORNIA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:OH, JONG DAE;CHIU, YI-JEN;LEE, SANG-HEE;AND OTHERS;SIGNING DATES FROM 20110330 TO 20110401;REEL/FRAME:026103/0004

MAFP

Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

Year of fee payment: 4